--- title: Pretraining Data Filtering for Open-Weight AI Safety description: EleutherAI presents 'Deep Ignorance', a research paper studying pretraining data filtering as a tamper-resistant safety mechanism for open-weight LLMs. Unlike... canonical: https://daily.dev/posts/pretraining-data-filtering-for-open-weight-ai-safety-urdd3ei6n twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Pretraining Data Filtering for Open-Weight AI Safety | daily.dev og:description: EleutherAI presents 'Deep Ignorance', a research paper studying pretraining data filtering as a tamper-resistant safety mechanism for open-weight LLMs. Unlike... og:url: https://daily.dev/posts/pretraining-data-filtering-for-open-weight-ai-safety-urdd3ei6n og:image: https://api.daily.dev/og/posts/urdD3ei6N.png og:image:alt: Pretraining Data Filtering for Open-Weight AI Safety og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Pretraining Data Filtering for Open-Weight AI Safety **[EleutherAI](https://daily.dev/sources/eleutherai)** · 12 min read · 0 upvotes · 0 comments ## Summary EleutherAI presents 'Deep Ignorance', a research paper studying pretraining data filtering as a tamper-resistant safety mechanism for open-weight LLMs. Unlike post-hoc safeguards (refusal training, input filters) that are easily bypassed via jailbreaks or fine-tuning, filtering dangerous knowledge from pretraining data prevents models from acquiring it in the first place. Using biorisk knowledge (WMDP-Bio benchmark) as a testbed, they trained 6.9B parameter models on 550B tokens with and without filtered data. Key findings: (1) filtering can reduce biorisk benchmark performance to near-random chance with minimal degradation to general capabilities; (2) filtered models remain noticeably safer even after adversarial fine-tuning on biorisk papers, unlike circuit-breaking approaches which collapse immediately; (3) however, filtered models can still leverage dangerous knowledge provided in-context (e.g., via RAG), suggesting filtering should be combined with other defenses. The pipeline uses a blocklist of ~6k terms plus a ModernBERT-Large classifier, processing 400M+ documents with under 1% additional compute cost. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: --- Tags: [#open-source](https://daily.dev/tags/open-source), [#llm](https://daily.dev/tags/llm) [View this post on daily.dev](https://daily.dev/posts/pretraining-data-filtering-for-open-weight-ai-safety-urdd3ei6n) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Pretraining Data Filtering for Open-Weight AI Safety","url":"https://daily.dev/posts/pretraining-data-filtering-for-open-weight-ai-safety-urdd3ei6n","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/pretraining-data-filtering-for-open-weight-ai-safety-urdd3ei6n"},"datePublished":"2026-08-02T07:04:32.129Z","dateModified":"2026-08-02T07:32:05.602Z","description":"EleutherAI presents 'Deep Ignorance', a research paper studying pretraining data filtering as a tamper-resistant safety mechanism for open-weight LLMs. Unlike...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/32fb1d53cbc564e23288bedbb9ac2fa6?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/32fb1d53cbc564e23288bedbb9ac2fa6?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"EleutherAI","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"EleutherAI","logo":"https://media.daily.dev/image/upload/s--aAcRyWGF--/f_auto,q_auto/v1785654263/logos/eleutherai","url":"https://daily.dev/sources/eleutherai"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/pretraining-data-filtering-for-open-weight-ai-safety-urdd3ei6n","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"open-source,llm","timeRequired":"PT12M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"EleutherAI","item":"https://daily.dev/sources/eleutherai"},{"@type":"ListItem","position":3,"name":"Pretraining Data Filtering for Open-Weight AI Safety"}]} ```